byt5-small-wikipron-eng-latn-us-broad-ft-au-broad

This model is a fine-tuned version of bookbot/byt5-small-wikipron-eng-latn-us-broad on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1622
• Per: 0.3093
• Gen Len: 16.2807

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0002
• train_batch_size: 128
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Per	Gen Len
0.4533	1.0	243	0.1681	0.3224	16.2423
0.1917	2.0	486	0.1618	0.3133	16.255
0.1726	3.0	729	0.1580	0.3126	16.2665
0.1591	4.0	972	0.1599	0.3126	16.2692
0.1506	5.0	1215	0.1583	0.3104	16.2694
0.1425	6.0	1458	0.1608	0.312	16.2782
0.1364	7.0	1701	0.1599	0.3104	16.2858
0.1308	8.0	1944	0.1600	0.3105	16.2755
0.1268	9.0	2187	0.1613	0.3105	16.2764
0.1239	10.0	2430	0.1622	0.3093	16.2807

Framework versions

• Transformers 4.28.0.dev0
• Pytorch 1.13.1+cu117
• Datasets 2.9.0
• Tokenizers 0.13.2